A Class G amplifier is a linear amplifier powered by multiple rail voltages, where load current is drawn through the output stage from the lowest rail that does not clip the output signal. Since this architecture reduces the average voltage drop across the amplifier's output stage, the result is improved efficiency without sacrificing the low-noise/high-linearity performance of a class AB amplifier.
Typically, the selection between supply rails is based on whether an output signal voltage of the amplifier rises above or falls below a threshold reference output voltage (VOREF). In order to maximize efficiency without distorting the output signal, the transition from a lower rail to a higher rail is typically made just before the output begins to clip. However, clipping levels are load dependent. An amplifier with a 1052 load will clip at a lower voltage than the same amplifier loaded with 100Ω.
If VOREF is fixed at a particular value, lightly loaded conditions may result in diminished efficiency due to early tripping. In contrast, heavy loads may pull the clipping voltage below VOREF, with the result that the amplifier never transitions to the higher rail.
In other applications, an input signal voltage of the amplifier may be monitored and the supply rail switched when the input voltage rises above or falls below a threshold input voltage (VIREF). Such applications often require that VIREF be chosen based upon the load impedance and the gain of the amplifier circuit so that VIREF corresponds to an appropriate output signal voltage.